Self-navigated subspace reconstruction for real-time MR imaging of the vocal tract

Abstract

Purpose: Real-time MRI offers a continuous and dynamic view of the object being imaged. Researchers have applied real-time MRI to speech production, which allows for the visualization of the vocal tract during speech.

Methods: This study proposed applying self-navigated subspace reconstruction for real-time vocal tract imaging. We performed experiments on a clinical 3 T MRI using standard RF coils and rapid acquisition. Additionally, 1000 frames were compressed during reconstruction to a few principal components, and iterative low-rank approximation was performed on compressed k-space, in conjunction with the orthogonal basis estimation for the subspace.

Results: The simulation study involving a 32-time acceleration showed that the proposed method produced a reasonably small root mean square error (RMSE) of 0.159, compared to 0.278 for sliding window reconstruction, 0.2527 for SToRM and 0.294 for low-rank reconstruction. The study also presented in vivo images of a typical sagittal image with a temporal resolution of 7 ms/frame or 21 ms/frame for the three-slice scan.

Conclusion: Our study presented a subspace reconstruction technique that does not require a navigator echo, which can be used for real-time MRI, particularly in speech imaging applications.